Re: [PATCH v4] gdb/gcore: interrupt all threads before generating the corefile

[Simon Marchi <simark@simark.ca>]

> @@ -131,6 +132,28 @@ gcore_command (const char *args, int from_tty)
>    if (!target_has_execution ())
>      noprocess ();
>  
> +  scoped_restore_current_thread restore_current_thread;
> +  scoped_disable_commit_resumed disable_commit_resume ("generating coredump");
> +  struct inferior *inf = current_inferior ();
> +  scoped_finish_thread_state finish_state (inf->process_target (),
> +					   ptid_t (inf->pid));
> +
> +  bool all_stop_was_running = false;
> +  if (exists_non_stop_target ())
> +    stop_all_threads ("generating coredump", inf);
> +  else
> +    {
> +      all_stop_was_running = any_thread_of_inferior (inf)->executing ();
I'm wondering what happens if you call this while the inferior is doing
an in-line step over, with one thread resumed and the others stopped.  I
think the others will have executing == false, so could you fall on one
that isn't executing?

What if we are in all-stop, but we resumed with schedlock on?

Note that I'd be fine with merging the patch without considering those
cases, since it's clearly an improvement already.

> diff --git a/gdb/testsuite/gdb.base/gcore-nonstop.exp b/gdb/testsuite/gdb.base/gcore-nonstop.exp
> new file mode 100644
> index 00000000000..4a240eb0872
> --- /dev/null
> +++ b/gdb/testsuite/gdb.base/gcore-nonstop.exp
> @@ -0,0 +1,77 @@
> +# Copyright 2022 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +# This testcase checks that when in non-stop mode with some threads running
> +# the gcore command can interrupt all threads, generate a core dump and
> +# restart threads as required.
> +

Please add a short comment explaining the goal of this test.

> +standard_testfile
> +
> +if { [prepare_for_testing "failed to prepare" \
> +	${testfile} ${srcfile} {pthreads debug}] } {
> +    return
> +}

prepare_for_testing starts GDB, but it gets restarted below anyway.  Use
build_executable to avoid the unnecessary spawn.

> +
> +save_vars { GDBFLAGS } {
> +    append GDBFLAGS " -ex \"set non-stop on\""
> +    clean_restart ${binfile}
> +}
> +
> +set lineno [gdb_get_line_number "Break here"]
> +if { ![runto $lineno] } {
> +    return
> +}
> +
> +# We should be stopped in thread 1 while thread 2 is running.
> +gdb_test_sequence "info threads" "info threads" {
> +    {Id\s+Target Id\s+Frame}
> +    {\*\s+1[^\n]*\n}
> +    {\s+2\s+[^\n]*\(running\)[^\n]*\n}
> +}
> +
> +set th1_pc ""
> +gdb_test_multiple "p/x \$pc" "fetch thread 1 PC" {
> +  -wrap -re "$::decimal = ($::hex)" {
> +    set th1_pc $expect_out(1,string)
> +    pass $gdb_test_name
> +  }

Could that use get_hexadecimal_valueof?

Simon